Adolescence is a critical period for the initiation and escalation of alcohol use and for the emergence of sex differences in alcohol use. The proposed research will test how changes in testosterone (T) during adolescence intersect with genetic influences on the development of alcohol use and related risk-taking behaviors. T levels increase dramatically from childhood through adolescence, particularly in males. Adolescent increases in T affect brain structure and function, including neural response to reward. Moreover, behavioral studies have found that individual differences in T predict alcohol use phenotypes, with stronger associations seen in boys than in girls. A largely independent stream of behavioral genetic research has established that genetic influences on substance use and related risk-taking behaviors increase over the course of adolescence. How these genetic influences intersect with hormonal changes during adolescence is unknown, as endocrine measures and behavioral genetic data have rarely been integrated. This gap is partly due to methodological challenges associated with measuring hormones in saliva: Single measures do not fully discriminate basal levels of hormones from state fluctuations (e.g., situational reactiviy, diurnal rhythm), while high-intensity repeated measurement is costly and burdensome to participants. The proposed research aims to overcome this barrier to research progress by using cutting-edge technology to measure accumulated T in hair. Hair T represents a 3-month hormonal accumulation and thus reflects chronic individual differences un- confounded with state fluctuations but sensitive to pubertal changes. We will examine the interplay between T and genetic influences on alcohol use in a sample of 500 twins (250 same-sex MZ and DZ pairs) ages 13-18. We will collect data on (1) T in saliva and hair, (2) reward sensitivity using a battery of in-lab behavioral tasks and self-report surveys, and (3) alcohol use using youth-report, parent-report, and school disciplinary records. This approach will address the following specific aims: (1) investigate the measurement of T in hair as a novel method that captures the underlying genetic signal better than salivary T, (2) examine T as an endophenotype that mediates genetic influences on alcohol use through its effects on reward sensitivity, and (3) examine T as a moderator of genetic influences on reward sensitivity and alcohol use (i.e., gene x hormone interaction). We hypothesize that accumulation of T in hair will represent a highly heritable endophenotype that both mediates and moderates genetic influences on alcohol use, with genetic risk being exacerbated in high T individuals. We also hypothesize that there will be sex differences in the gene > hormone > behavior links, with stronger associations evident in males than females, thus contributing to the emerging sex difference in alcohol use phenotypes. Given the novelty of measuring sex steroids in hair and the potential to illuminate genetic and epigenetic mechanisms underlying adolescent alcohol use, the project is both high-risk and high- reward, and is thus perfectly suited for the R21 mechanism.